Three different experimental murine animal models of human immunodeficiency virus type 1 (HIV-1) infection were evaluated in SCID mice for their technical feasibility, efficiency and distribution of successful lymphocyte transfer/engraftment. and susceptibility to detectable HIV-1 infection after reconstitution. Sources of reconstituted human cells that were studied included both fresh and frozen (cadaveric) bone marrow. fetal tissue, and peripheral blood. Based upon this analysis it was concluded that one model particularly well-suited to adaptation in the testing of putative antiretroviral agents was the SCID-hu-PBL model involving reconstitution of 8-12 week old SCID mice with fresh peripheral blood mononuclear cells obtained by lymphopheresis from normal donors. Reconstitution of SCID mice by intraperitoneal injection of harvested mononuclear cells led to successful and sustained transfer of lymphocytes that were optimally infectable 2-3 weeks after reconstitution. Human Iq antibody production was demonstrated in reconstituted mice as was trafficking of a small percentage of human lymphocytes to the murine spleen. In a series of experiments the optimal titer of infectious HIV-1 IIIB was determined using viral detection by both lymphocyte co-cultivation and polymerase chain reaction methods. Oral and intraperitoneally-administered zidovudine was found to demonstrate a dose-dependent protective effect in this model. Currently, this model is being adapted for the testing of other putative antiretroviral agents as well as for an assessment of the relative protection afforded by reconstitution with mononuclear cells derived from healthy gp-160-immunized donors.